Using Noncontact Measurement of Water Surface Dynamics to Estimate River Discharge

Estimating river discharge requires simultaneous measurement of velocity and flow depth. While surface velocities are relatively easy to measure using noncontact techniques, depth measurement usually requires physically intrusive instrumentation. This limits our capability to remotely monitor discharge in natural rivers subject to bed level variations. This work tests the potential to estimate the surface velocity, the water depth, the depth average velocity, and then discharge of a river using only a sequence of images of the dynamic water surface. The method is based on a comparison between the spatiotemporal Fourier spectra of the pixel intensities of these images and the theoretical dispersion relations of turbulence-generated surface fluctuations and gravity-capillary waves. The method is validated through the analysis of water surface videos obtained with fixed cameras from two river sections equipped with conventional discharge gauging. The applicability of the approach is demonstrated and the measurement uncertainties are quantified. The method is affected by two main sources of uncertainty: one derives from the estimation of the velocity index and the other from the obtainable resolution of the Fourier analysis. This resolution strongly controls the observation of depth and/or velocity variations in space and in time. The technique has advantages over current approaches: it has clear physical foundations; the equipment is low cost and is highly mobile; it does not need artificial tracers or physical equipment to measure depth; and it can directly provide estimates of the key flow parameters just from time series of images of the water surface.

Automated summary

This study examines the potential of estimating river surface velocity, water depth, depth average velocity, and discharge using only a sequence of dynamic water surface images. The method is based on comparing Fourier spectra of pixel intensities with theoretical dispersion relations, showing promising results for remote monitoring of river discharge.